Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

RMS, the most common soft tissue sarcoma of children, is an aggressive tumor of myogenic origin. Current treatment protocols include chemotherapy, surgery and radiation; although as with most cancers, metastatic disease is difficult to cure (Houghton et al., 1991). The development of pediatric RMS may be due, in part, to the failure of the normal differentiation program of myogenic precursors into myofibers. Cells arrested at a particular stage of development may retain growth and renewal properties resulting in the formation of sarcomas. The fetal development of endoderm to myofibers is controlled by a series of genes that act in a sequential fashion. MyoD, myogenin, myf5 and MRF4 have been identified as transcription factors necessary for the coordinate regulation of genes responsible for such differentiation (Braun et al., 1989a, 1989b, 1990; Edmonson and Olsen, 1989; Miner and Wold, 1990; Rhodes and Konieczny, 1990; Weintraub et al., 1991; Wright et al., 1989). Experiments with knockout mice have indicated the redundancy that some of these genes exhibit, for example, both myoD or myf5 mice develop normally, whereas myoD/myf5 double mutants die at birth due to the complete lack of skeletal muscle (Rudnicki et al., 1993). In addition, these mice lack expression not only of the mutant genes but also of myogenin and MRF4. The precise control of the expression of these transcription factors is not fully understood, but it is becoming apparent that feedback loops and the sequestration of partner proteins are important for regulation of gene expression (Dias et al., 1994; Weintraub, 1993).

RMS demonstrate overexpression of myoD, and because this protein is absent in mature muscle, myoD antibodies are used clinically to diagnose the disease (J. Jenkins, personal communication; Dias et al., 1990, 1992). Recently, the myoD promoter has been isolated and partly characterized, and a 258-bp enhancer element, which confers myogenic-specific expression in transgenic animals, was identified (Goldhamer et al., 1995; Faerman et al., 1995). If the myoD protein or appropriate transcription factors interact with this element, then cells overexpressing these proteins may allow enhanced expression of plasmid constructs containing these sequences.

To test the hypothesis that RMS-specific expression of potentially cytotoxic genes can be achieved, we designed a series of reporter constructs containing either the myogenin promoter alone or the myoD enhancer coupled to the SV40 promoter. Selectivity of expression was tested by transient transfection of plasmid DNA into a series of cells lines. After electroporation and analysis of cell extracts 48 hr after transfection, high level RMS-specific expression of reporter genes was detected from plasmids containing two copies of the myoD enhancer linked to the SV40 promoter. In addition, these transcriptional control elements, when ligated upstream of the HSVtk gene, conferred sensitivity to ganciclovir in RMS cell lines. Previously, investigators using this approach for the assessment of tumor-specific expression have selected individuals clones and performed survival assays on independent transfectants (Manome et al., 1994). Although this confirms the effectiveness of the promoter elements in this clone, it does not provide a true indication of the levels of expression in the entire population of cells. We have avoided this investigator bias by assaying pooled populations of transfected cells.

These data may provide the basis for development of viral vectors designed to specifically express genes that can be exploited to achieve RMS tumor-specific cell kill. If expression in normal tissues can be minimized, then the selective cytotoxicity may be conferred on tumor cells.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell lines and culture conditions. Table 1 indicates the source and references for the cell lines used in this study. All cell lines were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 2 mM glutamine except Molt-4, which contained 10% newborn calf serum.

Plasmids. The plasmids pgal-Basic, pgal-Promoter, pSEAP-Promoter and pSEAP-Control were purchased from Clontech (San Diego, CA). The control CAT reporter plasmid pCMVCAT, was obtained from Dr. L. Harris (St. Jude Children's Research Hospital). pSV--galactosidase was purchased from Promega (Madison, WI). pRc/RSV, encoding the neomycin resistance gene (neo), was obtained from InVitrogen (San Diego, CA). The plasmid pMYO1565LacZ containing the promoter elements and the first 18 nt of untranslated sequence of the myogenin gene linked to the Escherichia coli lacZ reporter was kindly provided by Dr. E. Olsen (M.D. Anderson, Houston, TX). The construction of the plasmids pgalM1, pgalM2, pMYOPROMTK, pMYOPROMKT and pCONTTK are described in this repory. Because multiple copies of repeat sequences can be unstable when grown in bacteria, all plasmids were isolated from E. coli STBL2 (Life Technologies, Gaithersburg, MD) grown at 30°C. For electroporation, plasmids were subject to large-scale preparation using the Wizard maxiprep resin (Promega), and DNA was further purified by size exclusion chromatography using UltrogelA-2 (IBF, Villeneuve la Garenne, France; Micard et al., 1985).

PCR isolation of the myoD enhancer. The amplification and construction of tandemly ligated copies of the 258-bp myoD enhancer has been described previously (Potter, 1996).

Cell extract preparation and Western analysis. Cell pellets were sonicated for 15 sec in a minimal volume of extraction buffer (50 mM Tris, pH 8.0, 300 mM NaCl, 0.5 mM dithiothreitol, 1 mM EDTA, 0.1% NP40, 1 mM sodium orthovanadate, 0.2 mM phenylmethylsulfonyl fluoride, 0.1 U/ml aprotinin, 10 µg/ml leupeptin), cooled on ice for 1 min and sonicated a further 5 sec. Sonication was performed using a Ultrasonic Homogenizer 4710 (Cole Palmer) with a microtip probe. After centrifugation at 70,000 × g for 30 min at 4°C, 30 µg of each sample was separated in 11% SDS-PAGE and transferred to Immobilon-P membrane (Millipore, Bedford, MA). Detection of specific proteins was accomplished using an ECL Kit (Amersham, Arlington Heights, IL) and exposure to Biomax MR film (Eastman Kodak, Rochester, NY). Monoclonal antibodies to myogenin and myoD were kindly provided by Drs. W. Wright (M.D. Anderson) and P. Houghton (St. Jude Children's Research Hospital), respectively. An anti-desmin monoclonal antibody was purchased from Dako (D33; Carpinteria, CA). An antibody to -tubulin (TUB2.1; ICN Biomedicals, Costa Mesa, CA) was used as a loading control in all analyses. To remove antibodies from the membrane to allow reprobing with subsequent reagents, filters were incubated in 62.5 mM Tris, pH 6.7, 2% SDS and 100 mM -mercaptoethanol at 50°C for 30 min.

Transient transfection. Electroporation conditions for each cell line were determined by performing a series of test transfections with the plasmid pSV--galactosidase with increasing voltages. Optimal conditions were considered to be those that generated high levels of -gal activity while not exceeding >95% cell kill. Typically, 1 × 10⁷ cells were electroporated in a total volume of 200 µl of phosphate-buffered saline using a BioRad electroporator (Hercules, CA) with a capacitance extender set at 960 µF. Electroporation voltages ranged from 150 to 240 V. After 48 hr, the medium was removed, and attached cells were harvested by trypsinization, washed three times with 50 mM Tris, pH 7.8, 1 mM EDTA and 150 mM NaCl and lysed by freeze thawing three times in 220 µl of lysis buffer (50 mM Tris, pH 7.8, 1 mM EDTA). After centrifugation at 14,000 × g for 10 min at 4°C, the supernatant was used for -gal and CAT assay. Each series of transfections included a positive (pSV--gal) and negative (pbgal-Basic) control plasmid. All transfections were performed in triplicate. Because transfection efficacy varied between each cell line, the transfection frequency of each electroporation was determined by cotransfection with pCMVCAT (cytomegalovirus promoter regulating expression of CAT).

-Gal assays. Enzyme activity was determined by the conversion of chlorophenol red--galactopyranoside (Boehringer-Mannheim, Indianapolis, IN) to chlorophenol red as described previously (Eustic et al., 1991). All assays included both positive and negative controls and reactions lacking cell extract. Data was expressed as A₅₇₀/mg/hr and corrected for transfection efficiency by assaying CAT activities of identical cell extracts.

CAT activity assay. CAT activities in heat-treated cell extracts (65°C for 10 min) were determined using the liquid scintillation diffusion assay (Neumann et al., 1987). Positive and negative controls were included in all assays. DPM were recorded every 10 to 15 min, and activity was determined per mg of total protein per hour from the linear portion of the curve.

Clonogenic cell survivals. To assess the sensitivity of cells expressing HSVtk to ganciclovir, 1 × 10⁷ cells of RD, Rh30L and Saos-2 were cotransfected by electroporation with 20 µg of test plasmid and 5 µg of pRc/RSV. The latter plasmid encodes the neomycin resistance gene and allows for the selection of transfectants by addition of G418 to the cell culture media. The plasmid pRc/RSV was included in the transfections to reduce the number of nontransfected cells in the subsequent clonogenic assays. Two days after electroporation, cells were trypsinized, plated at an appropriate density in 3.5-cm wells, transfectants selected by the addition of G418 to media and ganciclovir added at concentrations equivalent to an IC₅₀ for each untransfected cell line. Various combinations of drug scheduling were tried, but administration of both G418 and ganciclovir simultaneously, resulted in maximal cell kill in the clonogenic survival assays. Because the transfection efficiency varied among the three cell lines, 2 × 10⁴ cells were plated for RD and Rh30L and 3 × 10⁴ cells for Saos-2. After 7 days, foci were stained with crystal violet and quantified using an Artek automatic colony counter. For each transfection, data were expressed as a percentage of surviving colonies compared with untreated cells. Data generated from cells transfected with pMYOPROMKT and treated with ganciclovir was arbitrarily set to 100%. Clonogenic survival results were normalized to data generated from the identical treatment of cells transfected with the control plasmid (pMYOPROMKT) after exposure ganciclovir.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Western analysis of cell extracts. In an attempt to construct a RMS-specific expression system, we analyzed the expression of genes involved in the myogenic determinant lineage with an aim to use the promoter elements from these genes. Because monoclonal antibodies to myogenin and myoD were readily available, we examined their expression in a panel of human cell lines. Figure 1 indicates the results of western analyses of cell extracts when probed with anti-myogenin, anti-myoD, anti-desmin or anti--tubulin antibodies. As a positive control, the mouse myoblast line, C₂C₁₂, was included in the Western analysis. As can be seen, the RMS cell lines CT-TC, RD, Rh7, Rh28, Rh28 L-PAM, Rh30H, Rh30L, Rh36 and SCMC-MM-1-1T demonstrate high level expression of both myogenin and myoD. Rh18 demonstrates very low levels of myoD although it maintains substantial amounts of myogenin.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Analysis of expression of myogenin, myoD and desmin in various cell lines. Cell extracts were separated in 11% SDS-PAGE, and proteins were transferred to Immobilon-P membrane. Specific immunoreactivity to myoD, myogenin and desmin was determined by sequential incubation of the filter with monoclonal antibodies to the appropriate protein and detection by ECL. The membrane was probed with -tubulin to confirm equal loading of protein samples.

Construction of chimeric myoD enhancer/SV40 promoter reporter plasmids. As expression of myogenin and myoD was detected in the majority of bona fide RMS cell lines, we undertook a series of experiments to determine whether the appropriate enhancer and promoter elements from these genes could confer specific expression in these cell lines. Initial experiments with a reporter plasmid containing the proximal 1565 bp of the myogenin promoter (pMYO1565LacZ) showed expression of -gal in several different cell types, including non-RMS cell lines (data not shown). The recent demonstration of muscle-specific expression from the myoD promoter in transgenic mice and the isolation of a putative enhancer element (Goldhamer et al., 1995; Faerman et al., 1995) allowed us to design a series of reporter constructs to determine whether RMS-specific expression could be achieved using these regulatory sequences.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Diagrammatic representations of the plasmids used in this manuscript. A, Reporter plasmids. B, Plasmids used for transient clonogenic assays. For simplicity, common plasmid sequences (e.g., the ampicillin resistance gene) are not shown. The arrow associated with the HSVtk gene indicates the orientation of the coding sequence with respect to the promoter elements. SV40P, SV40 promoter; lacZ, E. coli lacZ gene; SV40pA, SV40 polyadenylation signal; SV40 Enh, SV40 enhancer; myoD Enh, myoD enhancer.

RMS-specific expression of -gal from myoD enhancer/SV40 promoter reporter plasmids. Cell lines were transiently transfected with reporter constructs containing either one or two copies of the myoD enhancer, pgalM1 and pgalM2, linked to the SV40 promoter. Figure 3 shows the expression of -gal from these plasmids after transfection into a variety of cell lines. Data is expressed as the fold increase activity over pgal-Promoter when corrected for transfection frequency. Results were obtained by dividing the test sample for each cell line with the control pgal-Promoter from the same series of transfections. Standard deviations are included for each data set.

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Promoter activity from the plasmids pbgalM1 and pbgalM2 in various cell lines after transient transfection. Cells were transfected with 20 µg of test plasmid and 5 µg of pCMVCAT. gal and CAT activity were determined after 48 hr using colorimetric and liquid scintillation diffusion assays, respectively. The number after the cell line name indicates the plasmid used for the transfection (e.g., CT-TC-1 represents CT-TC transfected with pgalM1 and CT-TC-2 represents CT-TC transfected with pgalM2). Activity is expressed as fold greater than the parent plasmid lacking any enhancer sequences (pgal-Promoter) The line on the ordinate axis indicates the activity of pgal-Promoter; hence, values above the line indicate increases in relative promoter activity and conversely, values below the line are reductions in relative promoter activity.

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Promoter activity from the plasmids pgalM1 and pgalM2 compared with pSV--gal in various cell lines after transient transfection. Cells were transfected with 20 µg of test plasmid and 5 µg of pCMVCAT. -Gal and CAT activity were determined after 48 hr using colorimetric and liquid scintillation diffusion assays, respectively. The number after the cell line name indicates the plasmid used for the transfection (e.g., CT-TC-1 represents CT-TC transfected with pgalM1 and CT-TC-2 represents CT-TC transfected with pgalM2). Activity is expressed as a percentage of pSV--gal expression after transfection into the same cell line.

Construction of plasmids expressing HSVtk under control of the SV40 promoter and myoD enhancer. To determine whether the SV40 promoter/myoD enhancer was sufficient to exact cell specific cytotoxicity, we placed these control elements upstream of the HSVtk coding sequence. The schematic for the construction is shown in figure 5. Briefly, the HSVtk coding sequence was isolated by PCR from pBRTK6 (McKnight, 1982) using Pfu polymerase and primers that create BglII restriction enzyme sites, 58 bp upstream of the ATG initiation codon and 12 bp downstream of the termination codon. After digestion with BglII, the 1.2-kb fragment was ligated into pSP64Cla (pSP64 containing ClaI sites flanking the multiple cloning sites; P. M. Potter, unpublished data). This allowed the removal of the tk coding sequence as a 1.2-kb HindIII/ClaI fragment. The secreted alkaline phosphatase (SEAP) protein coding sequence was removed from the plasmid pSEAP-Promoter by digestion with HindIII and ClaI and the HSVtk sequence ligated in to generate pPROMTK. This plasmid contains the SV40 promoter but lacks the myoD enhancer elements.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Construction of the plasmid pMYOPROMTK. Briefly, the SEAP coding sequence was removed from pSEAP-Promoter and replaced with the HSVtk gene. Two copies of the myoD enhancer tandemly arranged as a BglII/BamHI fragment were then ligated into the unique BglII site immediately upstream of the SV40 promoter. Similar plasmids containing the SV40 enhancer/SV40 promoter regulating expression of HSVtk and with the HSVtk gene in the incorrect orientation with respect to the promoter elements were also constructed. B, BamHI; Bg, BglII; C, ClaI; H, HindIII; tk, HSV thymidine kinase; MyoD Enh, myoD enhancer; SEAP, secreted alkaline phosphatase; SV40I, SV40 intron; SV40pA, SV40 polyadenylation signal; SV40P, SV40 promoter.

Clonogenic survival of cells transfected with pMYOPROMTK after treatment with ganciclovir. Before electroporation, the cytotoxicity of several cell lines to G418 and ganciclovir was assessed by clonogenic survivals. IC₉₉ values for G418 were 500, 500 and 600 µg/ml for RD, Rh30L and Saos-2, respectively. IC₅₀ values for ganciclovir for RD, Rh30L and Saos-2 cells were 425, 115 and 420 µM, respectively. After transfection with the test plasmid and pRc/RSV and selection with G418, sensitivity to ganciclovir was assessed. Figure 5 indicates the clonogenic survival of transfected cell lines to ganciclovir (IC₅₀) after electroporation with pMYOPROMTK, pMYOPROMKT and pCONTTK. As can be seen, all three cell lines demonstrate reduced clonogenic survival when transfected with the control plasmid, pCONTTK (fig. 6A), indicating that the SV40 promoter/enhancer is active in all cell types. However transfection with pMYOPROMTK results in selective RMS cell kill because cell survival is greatly reduced in RD and Rh30L, whereas little reduction in survival is seen in Saos-2 (fig. 6B). Because equitoxic doses of ganciclovir were used in the survival assay (IC₅₀), data are comparable between cell lines of inherently different sensitivity to this drug. Growth inhibition analyses of transiently transfected Rh30 cells treated with ganciclovir demonstrated a 30% drop in the IC₅₀ value for cells electroporated with pMYOPROMTK compared with pMYOPROMKT (data not shown), confirming the clonogenic survival results.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Clonogenic survival of transiently transfected cells after selection with G418 and treatment with ganciclovir. After transfection with either pMYOPROMTK, pMYOPROMKT or pCONTTK DNA, cells were allowed to recover for 48 hr and then plated into media containing ganciclovir and G418. Colonies were allowed to form over a period of 7 days for each cell line and quantified using an Artek automatic counter. A, Survival of cells after transfection with pCONTTK compared with non-drug-treated cells. B, Relative survival of cells after transfection with pMYOPROMTK. Data are normalized to the percentage of surviving colonies after transfection with the negative control plasmid pMYOPROMKT.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

An analysis of a series of cell lines has indicated that the expression of myoD, myogenin and desmin can be detected in bona fide RMS. With the exception of the American Type Culture Collection lines A204, A673 and Hs729T and the St. Jude-derived cell line Rh1, expression of all three markers was readily apparent by Western analysis. We suspect that the extended time in culture of the former cell lines may have resulted in the down-regulation of these myogenic genes. Rh1 was originally diagnosed as an embryonal RMS, but it is now thought to be a tumor of primitive neuroectodermal origin (PNET). Because Rh1 was isolated from the patient in 1985 before myoD and myogenin antibodies were available, the primary diagnosis was determined from histopathology using conventional reagents. A recent immunohistochemical study of tumor samples using frozen sections essentially confirmed that all RMS expressed myoD and that this gene product was undetectable in normal tissues, including both adult and fetal muscle (Dias et al., 1992).

We took advantage of the selective expression of myoD and myogenin in RMS and designed a series of mammalian expression vectors containing the SV40 promoter and either one or two copies of the myoD enhancer. Transient transfection into eight RMS cell lines confirmed the expression of reporter genes from these constructs. The levels of -gal expression in RMS increased with additional copies of the 258-bp enhancer element, suggesting that the appropriate regulatory factors that interact with these sequences were present within these cell lines. While small increases in promoter activity in the non-RMS lines were observed with multiple myoD enhancer fragments, these were not significantly different from those seen after electroporation with pgalM1 (fig. 3).

Desmin levels in Rh30H were significantly higher than in Rh30L (fig. 1) suggesting that the former cell line is more differentiated than the latter. Additionally, reduced reporter gene expression was observed after transfection of Rh30H cells with both pgalM1 and pgalM1 in comparison to Rh30L (fig. 3). These data suggest that prolonged culture of RMS cell lines may result in the loss of expression of myoD, myogenin or other myogenic regulatory proteins and strongly implicate the transcription factors in the control of gene expression from these constructs. Because the American Type Culture Collection cell lines A204, A673 and Hs729T have been cultured for extended time periods, this may explain the lack of myogenic markers observed in these cells (fig. 1). Indeed, myoD-mediated promoter activity was very low in A673, similar to the non-RMS cell lines (fig. 3), suggesting that the necessary transcriptional control proteins are absent in this cell line. We did not perform transfections of C2C12 or Rh28 L-PAM because the former cell line is derived from mouse myoblasts and the latter is a derivative of Rh28. Additionally, the human RMS line SCMC-MM-1-1T has growth arrested, and we have been unable to generate enough cells to perform the transient transfections.

While the level of myoD enhancer-mediated SV40 promoter activity was relatively high in RMS cell lines, the activity in comparison to the SV40 enhancer and promoter varied dramatically between cell lines (fig. 4). This is to be expected because the SV40 transcription cassette demonstrates different activities in a variety of cell lines. Many mammalian expression vectors have been designed containing these promoter regulatory elements, but the efficiency of expression after transfection is highly dependent on the cell type. Our data indicate that while expression from the myoD enhancer/SV40 promoter varied within the RMS cell lines, in two cases, the promoter activity exceeded that observed with the SV40 enhancer (RD and Rh7). These data indicate the efficacy of expression from the myoD enhancer/SV40 promoter in RMS cell lines.

The overexpression of myoD in RMS, apparently controlled at the level of transcription, has allowed the generation of a hybrid myoD enhancer/SV40 promoter with the potential to allow selective expression of cytotoxic genes in these cells. To confirm the effectiveness of this chimeric transcriptional control element, the HSVtk gene was ligated downstream of the transcriptional initiation site and cell lines transiently transfected by electroporation. Cytotoxicity to ganciclovir was assessed by clonogenic survival and demonstrated that the RMS cell lines RD and Rh30L were sensitive to this drug. In contrast, Saos-2, which does not express myoD or myogenin and showed very little myoD enhancer-mediated reporter gene expression, was insensitive to this agent. Levels of cell kill were similar in RD and Rh30L, consistent with the similar levels of promoter activity afforded by the myoD enhancer (fig. 3). Because the clonogenic assays were performed with transiently transfected cells, a broad range of intracellular HSVtk protein levels would be present, presumably due to clonal variation and level of gene expression. This may explain the lack of complete cell kill in these assays. Stable transfections to isolate cell lines expressing high levels of HSVtk could be performed. However, because variations in the levels of the myogenic transcription factors have been observed within the same cell lines (Rh30L and Rh30H), we adopted a procedure that would not arbitrarily select for high or low level of HSVtk-expressing clones. This eliminated any investigator bias that could have been introduced during the selection and analysis of individual transfectants.

Overall, these data imply that the selective expression of target genes by the myoD enhancer/SV40 promoter in RMS cell lines, as demonstrated by both reporter and transient clonogenic assays, may be translated into a potential approach to in vivo therapy. Selective expression of HSVtk within tumor cells would allow specific cell kill after treatment with ganciclovir. In addition, not every tumor cell would need to be transduced because the bystander effect would result in toxicity to neighboring cells. Because 30% of RMS recur locally, the application of a viral delivery vehicle containing the myoD enhancer/SV40 promoter/cytotoxic gene immediately after resection might reduce any residual tumor and prevent or delay relapse.

Similar approaches using the DF3/MUC1 promoter and the secretory leukoprotease inhibitor promoter have been pursued for the tumor-specific expression of HSVtk in breast cancer and cervical carcinoma cell lines, respectively (Chen et al., 1995; Manome et al., 1994; Wang et al., 1996). If efficient transduction of tumor cells can be achieved in vivo, the selective cell kill afforded by these constructs may provide an additional therapeutic modality. The specific expression of myogenic transcription factors in RMS may provide a route for the delivery of genotoxic agents to these cells. Because alveolar RMS metastasizes early, usually resulting in the presentation of children with disseminated disease at diagnosis, the 5-yr survival for these patients is very low (~20%). If delivery of cytotoxic genes can be achieved by adenovirus for example, then selective tumor cell kill may be afforded by the expression of HSVtk in combination with ganciclovir administration. Expression in normal tissues including skeletal muscle would be expected to be very low (since myoD is not expressed in adult muscle); hence, transduced nonmalignant cells would not be affected by therapy.

We are currently generating adenovirus containing the myoD enhancer and SV40 promoter to determine whether RMS-specific expression can be achieved in immunodeprived mice bearing human RMS xenografts.